Published on Web 03/19/2002

Rate-Equilibrium Relationships in Hydride Transfer

Reactions: The Role of Intrinsic Barriers
Ernst-Ulrich Würthwein,*,† Gabriele Lang,‡,§ Ludwig H. Schappele,‡,| and
Herbert Mayr*,‡
Contribution from the Institut für Organische Chemie der Westfälischen Wilhelms-UniVersität
Münster, Corrensstrasse 40, D-48149 Münster, Germany, and Department Chemie der
Ludwig-Maximilians-UniVersität München, Butenandtstrasse 5-13 (Haus F),
D-81377 München, Germany
Received September 10, 2001

Abstract: A literature survey on the kinetics of hydride abstractions from CH-groups by carbocations reveals
a general phenomenon: Variation of the hydride acceptor affects the rates of hydride transfer to a
considerably greater extent than an equal change of the thermodynamic driving force caused by variation
of the hydride donor. The origin of this relationship was investigated by quantum chemical calculations on
various levels of ab initio and DFT theory for the transfer of an allylic hydrogen from 1-mono- and 1,1-
disubstituted propenes (XYCdCH-CH3) to the 3-position of 1-mono- and 1,1-disubstituted allyl cations
(XYCdCH-CH2+). The discussion is based on the results of the MP2/6-31+G(d,p)//RHF/6-31+G(d,p)
calculations. Electron-releasing substituents X and Y in the hydride donors increase the exothermicity of
the reaction, while electron-releasing substituents in the hydride acceptors decrease exothermicity. In line
with Hammond’s postulate, increasing exothermicity shifts the transition states on the reaction coordinate
toward reactants, as revealed by the geometry parameters and the charge distribution in the activated
complexes. Independent of the location of the transition state on the reaction coordinate, a value of 0.72
is found for Hammond-Leffler’s R ) δ∆Gq/δ∆rG° when the hydride acceptor is varied, while R ) 0.28
when the hydride donor is varied. The value of R thus cannot be related with the position of the transition
state. Investigation of the degenerate reactions XYCdCH-CH3 + XYCdCH-CH2+ indicates that the
migrating hydrogen carries a partial positive charge in the transition state and that the intrinsic barriers
increase with increasing electron-releasing abilities of X and Y. Substituent variation in the donor thus
influences reaction enthalpy and intrinsic barriers in the opposite sense, while substituent variation in the
acceptor affects both terms in the same sense, in accord with the experimental findings. Marcus theory is
employed to treat these effects quantitatively.

Introduction correlation between structural effects on rates and equilibria of

It has long been known that there is a general correspondence
between rates and heats of chemical reactions. Within a series
of related reactions, very exothermic reactions usually take place
faster than less exothermic ones. Numerous quantitative treat-
ments of such relationships have been reported (eq 1),1,2 and
Dewar suggested referring to them collectively as the Bell-
Evans-Polanyi principle.3
∆Eq ) A + B∆H (1)
One of the consequences of the Bell-Evans-Polanyi prin-
ciple is the well-known Hammond postulate4 and the Ham-
mond-Leffler relationship1 (eq 2), which provides a quantitative
* To whom correspondence should be addressed. E.-U.W.: Fax (inter-
nat.), +49(0)251/83-39772; E-mail, wurthwe@uni-muenster.de. H.M.: Fax
(internat.), +49(0)89/2180-7719; E-mail, Herbert.Mayr@cup.uni-muenchen.de.
† Westfälische Wilhelms-Universität Münster.
‡ Ludwig-Maximilians-Universität München.
§ Current address: BASF AG, Polymer Laboratory, D-67056 Ludwig-
shafen, Germany.
| Current address: Ems-Dottikon AG, CH-5605 Dottikon, Switzerland.
4084 9 J. AM. CHEM. SOC. 2002, 124, 4084-4092
chemical reactions.
δ∆Gq ) Rδ∆G° (2)
δ∆G° expresses the change of the reaction free energy,
induced by structural variation, and the proportionality constant
R reflects the fraction of this effect that is observable in the
activation free energy ∆Gq. The magnitude of R has often been
used for localizing transition states, but it is well-known that
the assumption 0 < R < 1 and the association of small values
of R with early transition states, and of large values of R with
late transition states, is too simplistic.5 Such relationships have
extensively been investigated in proton-transfer reactions.
Changes of ∆G° due to variation of the base have been found
(1) Leffler, J. E.; Grunwald, E. Rates and Equilibria of Organic Reactions;
Wiley: New York, 1963.
(2) Chapman, N. B.; Shorter, J. AdVances in Linear Free Energy Relationships;
Plenum Press: London, 1972.
(3) Dewar, M. J. S.; Dougherty, R. C. The PMO Theory of Organic Chemistry;
Plenum Press: New York, 1975.
(4) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334-338.
10.1021/ja0121540 CCC: $22.00 © 2002 American Chemical Society
Relationships in Hydride Transfer Reactions ARTICLES

Table 1. Calculated (AM1) and Experimental Hydride Affinities (kcal mol-1) of Some Conjugated Carbenium Ions Derived from the
Corresponding Heats of Formation
AM1 gas-phase data
R−H ∆fH°(R−H) ∆fH°(R+) HIAa ∆fH°(R−H) ∆fH°(R+) HIAb
cycloheptatriene (1a) 38.2 210.4 257.8 44.6c 208.9d 199.0
1,4-cyclohexadiene (1b) 16.9 205.9 274.6 25.0c 207.1e 216.8
1,4-dihydronaphthalene (1c) 27.4 213.0 271.2 33f 210.5e 212
9,10-dihydroanthracene (1d) 38.4 221.1 268.3 38.2g 213.1e 209.6
xanthene (1e) 18.2 198.1 265.5
4-methyl-2,5-heptadiene (1f) 2.5 181.9 265.0
2,5-heptadiene (1g) 4.8 190.9 271.7
3-methyl-1,4-pentadiene (1h) 20.6 211.8 276.8
1,4-pentadiene (1i) 24.6 222.5 283.5 25.2g (220.4)d 229.9
3-propylcyclopentene (1j) -15.2 175.6 276.4
p-cymene (1k) -3.0 183.6 272.2 -7.4g 173.8d 215.9
allylbenzene (1l) 33.2 225.0 277.4

a ∆ H°(H-) ) 85.58 kcal mol-1 (AM1). b ∆ H°(H-) ) 34.7 kcal mol-1, from ref 10. c From ref 10. d From ref 11. e Calculated with ∆ H°(arene) from
f f f
ref 10, proton affinities for arenes from ref 12a, and ∆fH°(H+) ) 365.7 kcal mol-1 from ref 10. f Calculated from ∆fH°(1c) ) 20.13 kcal mol-1 in the
-1 g
condensed phase plus a vaporization enthalpy of 13 kcal mol as for 1,2,3,4-tetrahydronaphthalene, from ref 13. From ref 13.

to affect ∆Gq differently as identical changes of ∆G° due to

variation of the acid. Bernasconi’s principle of nonperfect
synchronization has been formulated to account for those
findings.6 The kinetic data reported in the preceding paper
indicated that structural variations of allylic and benzylic hydride
donors had remarkably small effects on hydride transfer rates,
while these rates were strongly affected by variation of the
hydride acceptors.7 It was the goal of this work to search for
the origin of these unexpected relationships.
Correlation between the Nucleophilicity Parameters N and
Hydride Ion Affinities. Because of the problems to compare
nucleophilic reagents of large structural variety with respect to
a single reference electrophile, we have introduced eq 3, which
allows one to compare nucleophilic reagents of widely varying
reactivities.8,9

log k(20 °C) ) s(N + E) (3)

where s is the nucleophile-specific slope parameter, N is the
nucleophilicity parameter, and E is the electrophilicity parameter.
Because s has been found to be close to 1 for several CH
hydride donors,7 the relative reactivities of different hydride
donors (∆ log k) can be substituted by ∆N, and the slope of a
correlation (2.303RT)N versus hydride ion affinity (HIA) of the
resulting carbocations corresponds to Bell-Evans-Polanyi’s
B (eq 1) or Hammond-Leffler’s R (eq 2). The low quality of
the correlation between N and calculated hydride affinities
(AM1, Table 1) as well as the small slope of this correlation
(Figure 1) confirm, in agreement with the conclusions drawn
(5) (a) Bordwell, F. G.; Boyle, W. J., Jr.; Hautala, J. A.; Yee, K. C. J. Am.
Chem. Soc. 1969, 91, 4002-4003. (b) Marcus, R. A. J. Am. Chem. Soc.
1969, 91, 7224-7225. (c) Pross, A. AdV. Phys. Org. Chem. 1977, 14,
69-132; see refs on pp 93-96. (d) Lewis, E. S.; Hu, D. D. J. Am. Chem.
Soc. 1984, 106, 3292-3296. (e) Wiseman, F.; Kestner, N. R. J. Phys. Chem.
1984, 88, 4354-4358. (f) Lewis, E. S. J. Phys. Chem. 1986, 90, 3756-
3759. (g) Yamataka, H.; Nagase, S. J. Org. Chem. 1988, 53, 3232-3238.
(h) Bunting, J. W.; Stefanidis, D. J. Am. Chem. Soc. 1989, 111, 5834-
5839. (i) Pross, A.; Shaik, S. New J. Chem. 1989, 13, 427-433. (j) Arnaut,
L. G. J. Phys. Org. Chem. 1991, 4, 726-745. (k) Guthrie, J. P. J. Am.
Chem. Soc. 1996, 118, 12886-12890. (l) Lee, W. T.; Masel, R. I. J. Phys.
Chem. A 1998, 102, 2332-2341. (m) Yamataka, H.; Mustanir; Mishima,
M. J. Am. Chem. Soc. 1999, 121, 10223-10224. (n) Shaik, S.; Shurki, A.
Angew. Chem. 1999, 111, 616-657; Angew. Chem., Int. Ed. 1999, 38, 586-
625. (o) Toteva, M. M.; Richard, J. P. J. Am. Chem. Soc. 2000, 122, 11073-
11083. (p) See also ref 28.
(6) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119-238 and references
therein.
(7) Mayr, H.; Lang, G.; Ofial, A. R. J. Am. Chem. Soc. 2002, 124, 4076-
4083.
Figure 1. Correlation between the hydride reactivity N of various
hydrocarbons (CH2Cl2, 20 °C) and the calculated (AM1) hydride ion
affinities HIA of the resulting carbenium ions (N ) -0.1608 HIA + 42.43;
r ) 0.6941).
in the preceding article, that hydride transfer rates are not
predominantly controlled by the stabilization of the carbocations
produced.
Disregarding the large scatter of the correlation in Figure 1,
one can derive from its slope that roughly 20% of the differences
in the thermodynamic driving force are found as differences in
the hydride transfer rates (2.303RT × 0.1608 ) 0.216). An even
smaller value (15%) is calculated, when the correlation between
N and the six experimentally available hydride affinities (Table
1) is considered. This finding sharply contrasts the situation in
SN1 reactions, where approximately 89% of the differences of
the heats of ionization in solution14 and 53% of the differences
of the gas-phase hydride affinities are observable in the transition
states of the ethanolysis reactions.15
(8) (a) Mayr, H.; Patz, M. Angew. Chem. 1994, 106, 990-1010; Angew. Chem.,
Int. Ed. Engl. 1994, 33, 938-955. (b) Mayr, H.; Kuhn, O.; Gotta, M. F.;
Patz, M. J. Phys. Org. Chem. 1998, 11, 642-654. (c) Mayr, H.; Patz, M.;
Gotta, M. F.; Ofial, A. R. Pure Appl. Chem. 1998, 70, 1993-2000. (d)
Mayr, H.; Bug, T.; Gotta, M. F.; Hering, N.; Irrgang, B.; Janker, B.; Kempf,
B.; Loos, R.; Ofial, A. R.; Remennikov, G.; Schimmel, H. J. Am. Chem.
Soc. 2001, 123, 9500-9512.
(9) Lucius, R.; Loos, R.; Mayr, H. Angew. Chem. 2002, 114, 97-102; Angew.
Chem., Int. Ed. 2002, 41, 91-95.

J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002 4085

ARTICLES
Figure 2. Correlation between the rates of hydride transfer (CH3CN, 23
°C) from various triarylmethanes to the tritylium ion and pKR+ of the
resulting carbenium ions (log k ) 0.263 pKR+ - 4.652, r ) 0.992).16
An ) p-H3CO-C6H4; Tol ) p-H3C-C6H4.
Figure 3. Correlation between the rates of hydride transfer (CH3CN, 23
°C) from cycloheptatriene (log k ) -0.61 pKR+ - 3.73, r ) 0.999) and
tris(p-anisyl)methane (log k ) -0.46 pKR+ - 7.33, r ) 0.983) to tritylium
ions and the corresponding pKR+ values.16 An ) p-H3CO-C6H4.
It may be argued that the small slope of the correlation in
Figure 1 is due to the fact that kinetic data in solution are
correlated with thermodynamic data referring to the gas phase.
The following examples show, however, that similar conclusions
can be drawn from rate-equilibrium relationships, where kinetic
and thermodynamic data refer to the solution.
McDonough16 found a good correlation between hydride
abstraction rates from p-substituted triarylmethanes and pKR+
of the tritylium ions produced during this reaction (Figure 2).
Because the rate and equilibrium constants correlated in Figure
2 refer to almost the same temperature, the slope of this corre-
lation (0.263) can directly be compared with the slope derived
from Figure 1. One again finds that only a small percentage
(26%, slope in Figure 2) of the change of the reaction free
energy caused by variation of the hydride donor is found in the
transition states of the hydride transfer reactions. In contrast,
(log k)/pKR+ correlations that refer to reaction series, in which
the hydride acceptors (tritylium16 or 9-arylfluorenylium17 ions)
are altered, show much higher slopes (Figures 3 and 4).
In all reaction series, we observe the same phenomenon: a
small percentage of δ∆G° (15-26%) is encountered in the
(10) NIST Chemistry WebBook (NIST Standard Reference Database Number
69-July 2001 Release): http://webbook.nist.gov.
(11) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.;
Mallard, W. G. J. Phys. Chem. Ref. Data, Suppl. 1 1988.
(12) (a) Aue, D. H.; Guidoni, M.; Betowski, L. D. Int. J. Mass Spectrom. 2000,
201, 283-295. (b) Aue, D. H. In Dicoordinated Carbocations; Rappoport,
Z., Stang, P. J., Eds.; Wiley: New York, 1997; Chapter 3.
(13) Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemical Data of Organic
Compounds, 2nd ed.; Chapman & Hall: London, 1986.
4086 J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002
Würthwein et al.
Figure 4. Correlation between the rates of hydride transfer (CF3COOH,
30 °C) from bis(p-anisyl)methane (log k ) -0.73 pKR+ - 6.02, r ) 0.995)
and triphenylmethane (log k ) -0.60 pKR+ - 7.40, r ) 0.993) to
9-arylfluorenylium ions and the corresponding pKR+ values.17
Scheme 1
transition states of the hydride transfer reactions when the
hydride donor is varied, while a high percentage of δ∆G° (46-
73%) is found in the transition states when the hydride acceptor
is varied. To rationalize this behavior, the transition structures
of the hydride transfer reactions were calculated by semiem-
pirical, ab initio, and DFT MO methods. Because the identifica-
tion of electronic effects is most straightforward when the steric
situation at the reaction centers is kept constant, we have selected
the model reaction shown in Scheme 1, which refers to hydride
transfer processes of widely differing driving forces, while the
centers of the reactions remain unchanged.
Quantum Chemical Calculations of Reactants and Transi-
tion Structures. Computational Methods. All compounds
(substituted allyl cations and the corresponding propenes), the
transition states for hydride transfer, and the molecular com-
plexes (van der Waals complexes) were calculated at various
levels of theory. First, all structures were optimized using the
semiempirical method AM118 of the program package
MOPAC9319 (no symmetry constraint was applied). Then,
geometry optimizations using the RHF/6-31+G(d,p) basis set
of the Gaussian 98 program package20 were performed.
Frequency calculations for all stationary points were carried out
to characterize them either as minima (NIMAG ) 0) or as
transition states (NIMAG ) 1). Zero point energies are given
(14) Arnett, E. M.; Petro, C.; Schleyer, P. v. R. J. Am. Chem. Soc. 1979, 111,
522-526.
(15) Mayr, H. In Cationic Polymerizations: Mechanisms, Synthesis, and
Applications; Matyjaszewski, K., Ed.; Marcel Dekker: New York, 1996;
Chapter 2.
(16) McDonough, L. M. Ph.D. Thesis, University of Washington, 1960.
(17) Bethell, D.; Hare, G. J.; Kearney, P. A. J. Chem. Soc., Perkin Trans. 2
1981, 684-691.
(18) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am.
Chem. Soc. 1985, 107, 3902-3909.
(19) Stewart, J. J. P. MOPAC93; Fujitsu Ltd. 1993.
Relationships in Hydride Transfer Reactions ARTICLES

Table 2. Hydride Affinities (kcal mol-1) of R+-H,Me and

R+-Me,Me Calculated on Various Levels of Theory Compared to
the Experimental Values
methoda R+−H,Me R+−Me,Me
AM1//AM1 295.21 287.24
RHF/6-31+G(d,p)//RHF/6-31+G(d,p) 267.54 258.37
MP2/6-31+G(d,p)//RHF/6-31+G(d,p) 287.37 278.58
B3LYP/6-31+G(d,p)//B3LYP/6-31+G(d,p) 278.11 268.36
G2MP2 259.55 250.16
G3MP2 247.37 237.83
experimental 238.9b 230.4b

a Ab initio and DFT hydride affinities refer to ∆E ) E + ZPE. bFrom

0 tot
ref 12b.

unscaled, since scaling factors of 0.92 (tested for some exam-

ples) give reaction energies and free energies (Scheme 1, Tables
3-8) which deviate less than 0.2 kcal mol-1 from the unscaled
values. For the calculation of hydride affinities, a scaling factor
of 0.92 was used.21 The energies discussed (E0) are the sum of
single point MP2/6-31+G(d,p)//RHF/6-31+G(d,p) calculations
and zero point corrections (0 K). Free energies (∆G°, 298.15
K, 1.0 atm) are the sum of the MP2 single point calculations Figure 5. Relative energies of the van der Waals complexes and transition
and the thermal corrections to Gibbs’ free energy as obtained states for the degenerate reaction of R+-H,Me with R-H,Me calculated
from the frequency calculations. For calibrating purposes some on various levels of theory. B3LYP gives a minimum for the hydrido bridged
of the systems were calculated employing the B3LYP/6-31+G- species (NIMAG ) 0) and is therefore marked in parentheses. For the RHF,
MP2, and B3LYP calculations, the 6-31+G(d,p) basis set was used.
(d,p) DFT method as well as G2MP222a and G3MP2.22b Only
the latter method yields hydride affinities close to the experi-
mental values (Table 2), while all other methods give values
which are 20-60 kcal mol-1 too high, probably because of the
problems with calculation of the energy of the hydride ion.
The activation energies for these hydride transfer reactions
also depend highly on the level of computation as indicated by
Figure 5. RHF methods (AM1, RHF/6-31+G(d,p)) greatly
overestimate the activation barriers since unusually large
correlation energy effects are observed in the transition states.
Fortunately, the MP2/6-31+G(d,p)//RHF/6-31+G(d,p) energies
do not differ significantly from the G3MP2 results, allowing
us to use this much less expensive method. However, it is well
known23 that the DFT method B3LYP/6-31+G(d,p) underes- Figure 6. Molecular structure of the van der Waals complex W(Me,Me+-
timates the activation barriers, resulting in hydrido-bridged Me,Me) (RHF/6-31+G(d,p)//RHF/6-31+G(d,p)).
minima for the less stabilized systems. Whereas the RHF/
6-31+G(d,p) transition structures generally relax to give van
der Waals complexes, some of the DFT transition states relax
with formation of a new CC-bond (attack of the allyl cation at
the CC-double bond of the substituted propene). The basis set
superposition error (BSSE) appears to be negligible since an
increase of the basis set (6-31+G(2d,p), 6-31+G(3d,2p)) did
not alter the relative energies significantly.
Figure 7. Molecular structure of transition state TS(Me,Me-Me,Me)+.
(20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. Bond lengths H-C1 ) H-C4: 1.2898 Å (RHF/6-31+G(d,p)//RHF/
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, 6-31+G(d,p)).
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; For the calculation of the reaction pathways, the coplanar
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; anti-arrangement of the bonds C1/C2 and C4/C5 (see Figure 7)
Rabuk, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J.
V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; was chosen as starting geometry to avoid attack of the 1,1-
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, dimethylallyl cation at the CC-double bond or the lone pairs of
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, X and Y. Successive reduction of the distance between the
M.; Replogle, E. S.; Pople, J. A. Gaussian 98; Gaussian, Inc.: Pittsburgh, allylic hydrogen (1-H) and the acceptor carbon C4 leads to the
PA, 1998.
(21) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502-16513. transition state. Intrinsic reaction path calculations (IRC) as well
(22) (a) Curtiss, L. A.; Raghavachari, K.; Pople, J. A. J. Chem. Phys. 1993, 98, as full optimizations starting from the slightly disturbed transi-
1293-1298. (b) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K.; Rassolov,
V.; Pople, J. A. J. Chem. Phys. 1999, 110, 4703-4709. tion state geometries led to molecular complexes of the van
(23) (a) Johnson, B. G.; Gonzalez, C. A.; Gill, P. M. W.; Pople, J. A. Chem. der Waals type which represent the true minima of these parts
Phys. Lett. 1994, 221, 100-108. (b) Csonka, G. I.; Johnson, B. G. Theor.
Chem. Acc. 1998, 99, 158-165. of the potential energy surfaces.

J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002 4087

ARTICLES Würthwein et al.

Table 3. Total Energies (au)a and Zero Point Energies (kcal mol-1) of the Hydride Donors R-X,Y and the Resulting Carbenium Ions
R+-X,Y and Calculated Reaction Energies (kcal mol-1) for the Hydride Transfer of Scheme 1 (MP2/6-31+G(d,p)//RHF/6-31+G(d,p))
X,Y Etot(R−X,Y) ZPE(R−X,Y) Etot(R+−X,Y) ZPE(R+−X,Y) ∆E0 ∆rG°(25°C)
Me,NH2 -211.90589 83.78 -211.07090 77.77 -33.22 -32.86
OH,OH -267.59849 60.22 -266.73547 53.67 -16.17 -15.80
Me,OH -231.73838 75.29 -230.87129 68.75 -13.60 -13.79
H,OH -192.54957 56.43 -191.66885 50.38 -4.56 -2.32
Me,Me -195.88350 90.71 -194.99280 82.95 0.00 0.00
H,Me -156.69735 72.00 -155.79273 64.31 8.79 8.72

a 1 au ) 627.51 kcal mol-1.

Table 4. Total Energies (au)a and Zero Point Energies (kcal Table 5. Total Energies (au)a and Zero Point Energies (kcal
mol-1) of the van der Waals Complexes W(X,Y+-X,Y) Obtained mol-1) of the Activated Complexes TS(Me,Me-X,Y)+ and
from R+-X,Y and R-X,Y (MP2/6-31+G(d,p)//RHF/6-31+G(d,p)) Activation Barriers ∆Eq and ∆Gq (kcal mol-1) for the Hydride
and the Corresponding Reaction Energies and Free Energies Transfer Reactions of Scheme 1
(MP2/6-31+G(d,p)//RHF/6-31+G(d,p))
X,Y Etot ZPE ∆E0 ∆rG°(25°C)
forward reaction backward reaction
Me,NH2 -423.01183 163.01 -20.53 -10.66
Me,OH -462.63284 145.01 -13.57 -4.25 X,Y Etot ZPE ∆Eqb ∆Gq (25 °C)c ∆Eqd ∆Gq (25 °C)c
Me,Me -390.88600 174.00 -5.74 0.53
OH,OH -462.59993 142.10 -6.50 3.65 9.67 19.45
H,Me -312.49923 136.74 -5.31 0.68
Me,OH -426.74071 157.55 -6.67 3.59 6.93 17.38
H,OH -387.54743 139.08 -3.49 8.46 1.08 10.77
a 1 au ) 627.51 kcal mol-1. Me,Me -390.88004 172.51 -3.49 6.50 -3.49 6.50
H,Me -351.68742 153.95 0.71 10.56 -8.08 1.84
Reaction Energies. According to Table 3, the reaction
a 1 au ) 627.51 kcal mol-1. b ∆Eq(forward) ) E (TS(Me,Me-X,Y)+)
energies for the hydride abstractions depicted in Scheme 1 range 0
- E0(R-X,Y) - E0(R+-Me,Me). c For calculation see footnotes d and e
from -33.2 kcal mol-1 for the formation of the highly stabilized of Table 10. d ∆Eq(backward) ) E0(TS(Me,Me-X,Y)+) - E0(R-Me,Me)
1-amino-1-methylallyl cation to +8.8 kcal mol-1 for the - E0(R+-X,Y).
formation of the less stabilized 1-methylallyl cation. Almost
the same values were calculated for the reaction free energies.
van der Waals Complexes. For the hydrocarbon systems
(X,Y ) Me and/or H), complexation energies of 5-6 kcal mol-1
are calculated (Table 4), in good agreement with experience.24
Systems with OH or NH2 substituents, however, give higher
complexation energies (13-21 kcal mol-1) due to hydrogen
bonding as shown for the reactions of R+-X,Y with R-X,Y
in Table 4. Particularly the loss of translational degrees of
freedom accounts for the 6-10 kcal mol-1 difference between
∆E0 and ∆rG°.
In the van der Waals complex W(Me,Me+-Me,Me) obtained
from R-Me,Me and R+-Me,Me (Figure 6), the plane of the
allyl cation is located perpendicular to the plane of the olefin,
with the shortest CC-distances of 3.79 and 3.93 Å between C4
Figure 8. Energy diagram for the reaction of R-Me,Me with R+-Me,Me.
of the cation and C2 and C3 of the olefin. The shortest CH-
distances between cation and olefin are 2.75 (4H-C2) and 2.89
structures which are energetically more favorable than the
Å (4H-C3) indicating substantial CH interaction. The distance
reactants (Table 5, Figure 8). The absolute barriers with respect
between the migrating 1H and the acceptor center C4 is
to the van der Waals complexes range from 0 to 26 kcal mol-1.
considerably longer (3.74 Å).
However, the highly exothermic reactions of R-Me,NH2 and
The van der Waals interactions are eliminated in the following R-NH2,NH2 with R+-Me,Me proceed to the products without
analyses by considering the activation energies with respect to barrier, and transition states could not be located for these
the sum of the starting materials and not relative to the energies reactions.
of the van der Waals complexes. As depicted for TS(Me,Me-Me,Me)+ in Figure 7, transition
Transition Structures for Hydride Transfer. The transition structures with almost linear arrangements of C1-H-C4 were
structures (Figure 7) have explicitly been optimized and calculated for all reactions listed in Table 5 (Table 6).
demonstrated to be saddle points on the energy hypersurface; Analogously, C-H-C angles of 180° and 160° were calculated
they possess one and only one negative force constant. These for (tBu-H-tBu)+ and (iPr-H-iPr)+, while this angle reduces
transition structures have rather small relative energies (∆Eq) to 133° and 106° for (Et-H-Et)+ and (Me-H-Me)+,
with respect to the separated reactants ranging from -8 to +10 respectively.25a,b
kcal mol-1. Table 6 and Figure 9 show that the highly exothermic reaction
As a consequence of the initial formation of van der Waals of R+-Me,Me with 1,1-dihydroxypropene (R-OH,OH) pro-
complexes, stationary points can also be found for transition
(25) (a) Frash, M. V.; Solkan, V. N.; Kazansky, V. B. J. Chem. Soc., Faraday
(24) (a) Kim, K. S.; Tarakeshwar, P.; Lee, J. Y. Chem. ReV. 2000, 100, 4145- Trans. 1997, 93, 515-520. (b) Boronat, M.; Viruela, P.; Corma, A. J. Phys.
4186. (b) Chałasiński, G.; Szczȩśniak, M. M. Chem. ReV. 2000, 100, 4227- Chem. B 1997, 101, 10069-10074. (c) Boronat, M.; Viruela, P.; Corma,
4252. A. J. Phys. Chem. B 1999, 103, 7809-7821.

4088 J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002

Relationships in Hydride Transfer Reactions ARTICLES

Table 6. Bond Angles and Bond Lengths in the Transition States

of the Hydride Transfer Reactions of Scheme 1
(RHF/6-31+G(d,p))
X,Y ∆E0, kcal mol-1 β(C1−H−C4) d(H−C1), Å d(H−C4), Å
OH,OH -16.17 177.87° 1.2429 1.3579
Me,OH -13.60 176.45° 1.2464 1.3633
H,OH -4.56 178.42° 1.2764 1.3138
Me,Me 0.00 179.82° 1.2898 1.2898
H,Me 8.79 177.44° 1.3463 1.2413

Figure 11. Brønsted plot for the forward reaction of Scheme 1 (variation
of the hydride donor, MP2/6-31+G(d,p)//RHF/6-31+G(d,p), energies in
kcal mol-1).

Figure 9. CH-bond lengths in the transition states of the hydride transfer

reactions of Scheme 1 (RHF/6-31+G(d,p)).

Figure 12. Brønsted plot for the backward reaction of Scheme 1 (variation
of the hydride acceptor, MP2/6-31+G(d,p)//RHF/6-31+G(d,p), energies in
kcal mol-1).

Scheme 2

Supporting Information), when one considers the sum of the

Figure 10. Distribution of charges q (in multiples of electron charge) in
the transition states of the hydride transfer reactions of Scheme 1 (RHF/
6-31+G(d,p), Mulliken analysis).
ceeds through an early transition state closely resembling the
reactants, while the endothermic reaction of R+-Me,Me with
R-H,Me proceeds through a late transition state in which the
bond being formed (dH-C4 ) 1.24 Å) is already shorter than
the bond being broken (dH-C1 ) 1.35 Å). In the symmetrical
transition state of the thermoneutral reaction (X,Y ) Me), the
equal H-C1/C4 distances are 1.29 Å. The computational results
thus are in perfect agreement with the expectations based on
Hammond’s postulate.26
The change from early to late transition states on the way
from exothermic to endothermic reactions can also be seen in
the charge distribution of the transition states (Mulliken analysis
from the RHF/6-31+G(d,p)-optimizations, see Table S6 in the
charges at the donor and the acceptor fragments (Figure 10) as
indication for the progress of the reaction. Again in agree-
ment with Hammond’s postulate, the positive charge is pre-
dominantly localized at the acceptor fragment in the early
transition state of the reaction of R+-Me,Me with R-OH,OH,
while more charge is localized at the donor fragment in the
late transition state of the corresponding reaction with R-H,Me
(Figure 10).
Let us now look at the relationship between ∆Gq and ∆rG°
which corresponds to Leffler-Hammond’s R as defined by eq
2. Figure 11 shows a fairly linear plot with a slope of 0.28. If
R were taken as a measure for the position of the transition
state, one would derive an early transition state for the whole
reaction series, in marked contrast to the conclusions drawn from
Figures 9 and 10.
Necessarily, for the reverse reaction (Figure 12) a high value
of R is derived (R ) 1 - 0.28 ) 0.72) which would be
indicative of a late transition state according to the Leffler-
Hammond postulate.

J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002 4089

ARTICLES Würthwein et al.

Table 7. Total Energies of the Transition States of the Table 9. Hydride Affinities of the Allyl Cations R+-X,Y and
Degenerate Reactions, Their Zero Point Energies, Bond Lengths Intrinsic Barriers of the Degenerate Reactions in Scheme 2 (in
to the Migrating Hydrogen, Bond Angles, and NBO Charges at the kcal mol-1; According to MP2/6-31+G(d,p)//RHF/6-31+G(d,p))
Migrating Hydrogen in the Transition States TS(X,Y-X,Y)+
(MP2/6-31+G(d,p)//RHF/6-31+G(d,p)) X,Y HIA (G3MP2) ∆Eq (∆G0q)ii

ZPE, d(H−C1) ) NBO

Me,NH2 204.24 5.30 14.95
OH,OH 220.60 -0.62 10.76
X,Y Etot, aua kcal mol-1 d(H−C4), Å β(C1−H−C4) chargesb
Me,OH 224.02 1.81 12.43
Me,NH2 -422.96419 158.95 1.3142 176.63° +0.151 H,OH 233.69 0.72 12.47
OH,OH -534.33192 112.00 1.3066 172.26° +0.153 Me,Me 237.83 -3.49 6.50
Me,OH -462.60465 142.70 1.3044 175.24° +0.147 H,Me 247.37 -5.14 4.71
H,OH -384.21550 105.71 1.2981 178.63° +0.145
Me,Me -390.88004 172.51 1.2898 179.82° +0.144
H,Me -312.49662 135.27 1.2847c 174.66°d +0.142 corresponding degenerate reactions (eq 5).27,28
a 1 au ) 627.51 kcal mol-1. b According to ref 30. c d(H-C1) ) d(H-
C4) ) 1.2856 Å at MP2(Full)/6-31G*//MP2(Full)/6-31G*. d β(C1-H-C4) (∆G0q)ij ) 0.5((∆G0q)ii + (∆G0q)jj) (5)
) 178.65° at MP2(Full)/6-31G*//MP2(Full)/6-31G*.

Table 8. Selected Bond Lengths (Å) of the Carbon Skeleton of To derive the intrinsic barriers according to eq 5, we have
the Hydride Donors R-X,Y, of the Hydride Acceptors R+-X,Y, performed ab initio calculations on the degenerate hydride
and of the Transition States TS(X,Y-X,Y)+ of the Degenerate transfer reactions shown in Scheme 2.
Reactions in Scheme 2 According to
MP2/6-31+G(d,p)//RHF/6-31+G(d,p) As shown above for the nondegenerate reactions (Table 6),
the arrangement C1-H-C4 is also almost linear in the tran-
R−X,Y R+−X,Y TS(X,Y−X,Y)+
sition structures of the degenerate reactions (Table 7), and the
X,Y d(C1−C2) d(C2−C3) d(C4−C5) d(C5−C6) d(C1−C2) d(C2−C3)
distance H-C1/4 decreases noticeably as the stabilization of
Me,NH2 1.5052 1.3293 1.3302 1.4544 1.4261 1.3722 the allyl cations is reduced. The increasing tightness of the
OH,OH 1.5019 1.3161 1.3340 1.4370 1.4280 1.3531
Me,OH 1.5036 1.3231 1.3382 1.4313 1.4340 1.3540 bonding to the migrating hydrogen in the transition structures
H,OH 1.5031 1.3175 1.3448 1.4103 1.4383 1.3439 TS(X,Y-X,Y+) is accompanied by an elongation of the
Me,Me 1.5051 1.3277 1.3496 1.4175 1.4484 1.3478 C1-C2 ()C4-C5) bond and a shortening of the C2-C3
H,Mea 1.5032 1.3227 1.3604 1.3927 1.4523 1.3379 ()C5-C6) bond, that is, an increasing bond alternation from
a At MP2(Full)6-31G*//MP2(Full)/6-31G* the following bond lengths top to bottom of Table 8.
(Å) were obtained:1.4975, 1.3385, 1.3714, 1.3967, 1.4375, 1.3591. Table 9 shows the decrease of the intrinsic barriers with
increasing hydride affinities of the corresponding allyl cations;
In agreement with the experimental results discussed in the that is, the intrinsic barriers decrease with decreasing stabiliza-
beginning, the calculations also lead to the conclusion that donor tion of the carbocations involved. Analogously, the tendency
variation has a small influence on the activation free energy (R to yield hydrido-bridged carbocations [R-H-R]+ was calcu-
) 0.28, Figure 11), while acceptor variation has a large influence lated to increase in the series tBu < iPr < Et < Me.25 In accord
(R ) 0.72, Figure 12). with these findings, NMR studies in superacidic solution
The fact that Leffler-Hammond’s R cannot be related to the revealed hydrido-bridging in cycloalkyl cations to be more
location of the transition state is best illustrated by the degenerate pronounced in secondary than in tertiary carbocations.29
reaction of R-Me,Me with R+-Me,Me. By looking at this The intrinsic barriers (∆G0q)ii of Table 9 can now be com-
reaction as a representative of the series in Figure 11 one would bined with the reaction free energies ∆rG° in Table 3 for cal-
derive an early transition state, whereas one would derive a late culating ∆Gq for the forward and backward reactions of
transition state for the same reaction when one looks at it as Scheme 1 using the Marcus formalism (eqs 4 and 5). It can
part of the reaction series of Figure 12. How can one explain be seen in Table 10 that the activation free energies derived
the different magnitude of R as one analyzes Scheme 1 in the from Marcus theory closely agree with the activation free
forward or backward sense? energies directly calculated by MP2. The different slopes of
Marcus Treatment of the Quantum Chemical Results. An the plots in Figures 11 and 12 thus find a straightforward
explanation for these relationships is provided by Marcus theory explanation.
which derives the activation free energy of a reaction ∆Gq from When 1,1-dihydroxy-propene (R-OH,OH) is replaced by
its intrinsic barrier ∆G0q and the reaction free energy ∆rG° (eq 2-butene (R-H,Me), ∆rG° increases by 24.5 kcal mol-1 (Figure
4).27 11), the same amount that ∆rG° of the backward reactions is
reduced (Figure 12). In both reaction series the intrinsic barrier
∆Gq ) ∆G0q + 0.5∆rG° + ((∆rG°)2/16∆G0q) (4)
(28) (a) Kreevoy, M. M.; Lee, I.-S. H. J. Am. Chem. Soc. 1984, 106, 2550-
2553. (b) Kreevoy, M. M.; Ostović, D.; Lee, I.-S. H.; Binder, D. A.; King,
The work term of the Marcus equation can be neglected in G. W. J. Am. Chem. Soc. 1988, 110, 524-530. (c) Lee, I.-S. H.; Ostović,
ion-molecule reactions, because Coulomb interactions are small D.; Kreevoy, M. J. Am. Chem. Soc. 1988, 110, 3989-3993. (d) Kim, D.;
Lee, I.-S. H.; Kreevoy, M. M. J. Am. Chem. Soc. 1990, 112, 1889-1894.
in these cases. While the intrinsic barrier may directly be (e) Kreevoy, M. M.; Kotchevar, A. T. J. Am. Chem. Soc. 1990, 112, 3579-
measurable for degenerate reactions (∆Gq ) ∆G0q), ∆G0q is 3583. (f) Kim, Y.; Truhlar, D. G.; Kreevoy, M. M. J. Am. Chem. Soc.
1991, 113, 7837-7847. (g) Lee, I.-S. H.; Jeoung, E. H.; Kreevoy, M. M.
not directly accessible for nondegenerate reactions and is J. Am. Chem. Soc. 1997, 119, 2722-2728. (h) Lee, I.-S. H.; Jeoung, E. H.
calculated as the arithmetic mean of the intrinsic barriers of the J. Org. Chem. 1998, 63, 7275-7279. (i) Lee, I.-S. H.; Jeoung, E. H.;
Kreevoy, M. M. J. Am. Chem. Soc. 2001, 123, 7492-7496.
(29) (a) Kirchen, R. P.; Sorensen, T. S. J. Am. Chem. Soc. 1979, 101, 3240-
(26) Analogous changes of the transition structures have been calculated for 3243. (b) Kirchen, R. P.; Ranganayakulu, K.; Rauk, A.; Singh, B. P.;
hydrogen radical abstractions from Me-H, Et-H, iPr-H, and tBu-H by Sorensen, T. S. J. Am. Chem. Soc. 1981, 103, 588-596. (c) Sorensen, T.
H: ref 5g. S. In Stable Carbocation Chemistry; Prakash, G. K. S., Schleyer, P. v. R.,
(27) Marcus, R. A. J. Phys. Chem. 1968, 72, 891-899. Eds.; Wiley: New York, 1997; pp 75-136.

4090 J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002

Relationships in Hydride Transfer Reactions ARTICLES

Table 10. Indirectly (Marcus) and Directly Calculated Activation

Free Energies (kcal mol-1) for the Reaction of R+-Me,Me with
R-X,Y (Scheme 1)
∆Gq(forward reaction) ∆Gq(backward reaction)
X,Y ∆r G°a ∆G0qb Marcusc MP2d Marcusc MP2e
OH,OH -15.80 8.63 2.54 3.65 18.34 19.45
Me,OH -13.79 9.47 3.83 3.59 17.62 17.38
H,OH -2.32 9.49 8.36 8.46 10.68 10.77
Me,Me 0.00 6.50 6.50 6.50 6.50 6.50
H,Me 8.72 5.61 10.82 10.56 2.10 1.84

a See data in Table 3. b Intrinsic barriers ∆G q ) (∆G q) were calculated

0 0 ij
according to eq 5 from data in Table 9. c From ∆rG° and ∆G0q (columns 2
and 3 of this table) according to eq 4. ∆G ) G298(TS(Me,Me-X,Y)+)
d q
- G298(R+-Me,Me) - G298(R-X,Y). e ∆Gq ) G298(TS(Me,Me-X,Y)+)
- G298(R-Me,Me) - G298(R+-X,Y).
decreases by 3 kcal mol-1, as one moves from R-OH,OH to
R-H,Me. The moderate increase of the activation free energy
when going from left to right in Figure 11 thus is due to the
fact that the contribution of the increasing reaction enthalpy is
partially compensated by the decreasing intrinsic barrier. In
contrast, reaction free energy and intrinsic barrier are altered in
the same sense when moving from X,Y ) OH,OH to X,Y )
H,Me in the backward reaction of Scheme 1, and thus are
responsible for the high slope of the correlation in Figure 12.
It can thus be concluded that the strong dependence of hydride
transfer rates on variation of the hydride acceptor is due to the
fact that structural effects that increase the thermodynamic
driving force lower the intrinsic barrier. The small dependence
of hydride transfer rates on variation of the hydride donor, on
the other hand, is due to the fact that factors that increase the
thermodynamic driving force raise the intrinsic barrier and thus
lead to a partial cancellation of both effects.
The variability of intrinsic barriers within reaction series has
previously been demonstrated for hydride transfer reactions
between NAD+ analogues by Kreevoy et al.28 We can now
compare the results of Kreevoy’s analysis with the calculated
transition structures for hydride transfer reactions. Kreevoy and
Lee28a showed that structural changes in either Ai+ or AjH that
have similar effects on Kij of eq 6 can have different effects on
intrinsic reactivity and, consequently, on kij.
Ai+ + AjH h AiH + Aj+
If the in-flight hydrogen carries a full negative charge (q )
-1), the differential quotient R ) δ(ln kij)/δ(ln Kij) should be
0 when Ai+ is varied, and one would expect R ) 1 when AjH
is varied. Only if q ) 0 should variations in Kij have equal
effects on kij independent of which reactant (Ai+ or AjH) is
changed.
Application of this qualitative analysis on the reaction series
shown in Figure 11 (R ) 0.28) and Figure 12 (R ) 0.72) leads
to the conclusion that the in-flight hydrogen carries a positive
partial charge. It has been suggested28a that the charge q of the
in-flight H, which is related to the tightness parameter τ by eq
7, can experimentally be determined as the slope of a plot of ln
kii against ln Kij.
q ) τ - 1 ) δ(ln kii)/δ(ln Kij)
Replacement of ln kii and ln Kij by the corresponding free
energy terms listed in Tables 3 and 9 leads to the correlation
shown in Figure 13, which indicates a positive partial charge
Figure 13. Correlation between the intrinsic barriers and the Gibbs reaction
energies of a series of hydride acceptors toward a reference hydride donor
[(∆G0q)ii ) 0.2256∆rG° + 8.1961, r ) 0.8450, n ) 6].
of q ) +0.2 on the in-flight hydrogen. In accord with these
conclusions a positive charge (q ) +0.15) on the migrating
hydrogen is found in the calculated NBO charges30 (Table 7).
These conclusions are in agreement with the results of our
experimental investigations7 and the experimental data presented
in Figures 1-4. In contrast, a partially hydridic character has
been derived for the in-flight hydrogen in hydride transfer
reactions between NAD+ analogues.28 The difference may be
a result of the lower hydride affinities of pyridinium ions as
compared to those of the carbocations discussed above.
The less tight transition states of hydride transfer reactions
between pyridinium derivatives can directly be seen in the
C-H-C bond lengths of the hydrido bridge.31 While Table 7
shows an elongation from 1.28 to 1.31 Å with increasing
stabilization of the allyl cation, a value of 1.36 Å has been
calculated (6-31G*) for the corresponding hydrido bridge
between pyridinium ions.31,32 Experiments and quantum chemi-
cal calculations thus indicate significant differences in hydride
transfer processes between typical carbocations and pyridinium
ions.
Conclusions
Hydride transfer reactions between typical carbocations
(6) proceed through tight transition states in which the migrating
hydrogen carries a partial positive charge. As a consequence,
R ) δ(ln kij)/δ(ln Kij) is smaller than 0.5 when the hydride donor
is varied and larger than 0.5 when the acceptor is varied.
Computational and experimental results indicate the decrease
of the intrinsic barriers with decreasing stabilization of the
corresponding carbocations and the inappropriateness of R to
locate the position of the transition states on the reaction
coordinate.
Acknowledgment. This and the preceding paper are dedicated
to Professor George A. Olah on the occasion of his 75th
(30) Weinhold, F. Natural Bond Orbital Methods. In Encyclopedia of Compu-
tational Chemistry; Schleyer, P. v. R., Allinger, N. L., Clark, T., Gasteiger,
J., Kollman, P. A., Schäfer, H. F., III, Schreiner, P. R., Eds.; Wiley:
Chichester, 1998; Vol. 3, pp 1792-1811.
(31) Wu, Y.-D.; Lai, D. K. W.; Houk, K. N. J. Am. Chem. Soc. 1995, 117,
(7) 4100-4108.
(32) Though a general relationship between bond lengths and bond orders has
been suggested by Pauling (Pauling, L. J. Am. Chem. Soc. 1947, 69, 542-
553), it has been reported that different parameters have to be employed
for different types of bonds: (a) Houk, K. N.; Gustavson, S. M.; Black, K.
A. J. Am. Chem. Soc. 1992, 114, 8565-8572. (b) Zipse, H.; Wang, L.-H.;
Houk, K. N. Liebigs Ann. 1996, 1511-1522.

J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002 4091

ARTICLES
birthday. Financial support by the Deutsche Forschungsgemein-
schaft (SFB-424, Münster; Ma 673/18-1, München) and the
Fonds der Chemischen Industrie is gratefully acknowledged.
We thank Prof. S. Grimme (Münster), Prof. H. Zipse (München),
and Prof. M. Kreevoy (Minneapolis) for helpful suggestions,
and Prof. D. Aue (Santa Barbara) for his help in selecting
appropriate gas-phase data for carbocations.
4092 J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002
Würthwein et al.
Supporting Information Available: Kinetic data from ref 16
and details of the calculations (energies of AM1, ab initio, and
DFT calculations); Gaussian archive entries (MP2/6-31+G(d,p)//
RHF/6-31+G(d,p)) (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA0121540